[05:06:35] Noob question: If I change a use a hex editor and change the code from lowercase to uppercase, would the software stop functioning correctly? http://www.moeen.com/blog/PostsImagesFiles/1f349d8d5302_F8FD/image_3.png [05:35:34] Godfarter: this channel is for questions about wikimedia [05:40:19] true that, but I am pretty sure some of the people here have the knowledge required to answer my question, which would be very grateful for and it would make me have warm and fuzzy feelings about the Wikipedia community which in turn would make me edit more and stick around longer. [05:41:43] Do you know the name of another channel where my question would be more appropriate? [05:41:48] sorry, IRC noob [06:35:14] oh network http://ganglia.wikimedia.org/latest/?r=day&cs=&ce=&m=mem_report&s=descending&c=Glusterfs+cluster+pmtpa&h=&host_regex=&max_graphs=0&tab=m&vn=&sh=1&z=small&hc=4 [06:58:14] thedj: is the broken thumb of https://commons.wikimedia.org/wiki/File:Copyright.svg history tracked somewhere? [08:35:58] Nemo_bis: i don't see it as broken [10:43:45] thedj[work]: the previous version [10:44:21] it was broken since September at least, perhaps March (last rsvg update according to the category note by mah) [12:58:48] Gerrit 503-ing heavily the past few hous. [12:58:48] hours [13:23:29] <^demon> siebrand: Do you know who was configuring ULS? There was an uncommitted change on fenari. [13:23:53] ^demon: no. Probably the same person that configured Wikidata. [13:24:15] <^demon> Oh, ok. Thought it might've been you guys :) [13:28:26] ^demon: I'd guess Reedy [13:28:39] <^demon> Yeah, I got it now :) [13:28:42] <^demon> Thanks [13:31:18] <^demon> siebrand: Also, your 503 problem...I'm seeing a new exception that I've not seen before. http://p.defau.lt/?JE1wRtsxtwuqu4FADPeUXQ [13:31:27] <^demon> I think this might be related. [13:31:42] <^demon> Oh, crap. [13:31:48] <^demon> Gerrit's hitting an exception and restarting. [13:32:36] I saw one too [13:33:31] <^demon> Likely many people did. [13:33:34] <^demon> Gerrit restarted itself. [13:38:05] <^demon> !log flushed gerrit's account_byemail cache, error log was complaining [13:38:18] Logged the message, Master [13:41:12] ^demon, is it possible to change the owner of a change? [13:41:38] <^demon> No, it's not. [13:41:48] <^demon> There's some unfinished work pending review upstream for it, but it's not done [14:39:01] notpeter: hi. here? [17:02:25] !log Applied Wikibase/lib/sql/WikibaseLib.sql on wikidatawiki [17:02:40] Logged the message, Master [17:05:13] aharoni: hey, sure I can take a look [17:20:13] !log Running \Wikibase\Utils::insertDefaultSites(); on wikidatawiki [17:20:26] Logged the message, Master [17:54:28] !log Ran patch-job_token.sql on wikidatawiki [17:54:41] Logged the message, Master [18:00:53] RoanKattouw: Fancy helping with a CORS issue on wikidata? [18:00:58] GET http://bits.wikimedia.org/static-1.21wmf3/extensions/UniversalLanguageSelector/i18n/en-gb.json 404 (Not Found) load.php:128 [18:00:58] XMLHttpRequest cannot load http://bits.wikimedia.org/static-1.21wmf3/extensions/UniversalLanguageSelector/i18n/en-gb.json. Origin http://www.wikidata.org is not allowed by Access-Control-Allow-Origin. [18:01:20] You +1'd the addition of json files on friday, which mutante merged, and you just graceful-ed for... [18:02:45] OH [18:02:50] just found a possible issue [18:13:44] RoanKattouw and James_F are arguing about how much Reedy needs to know about what RoanKattouw is doing :) [18:13:52] haha [18:13:59] Reedy: So yeah I'm pushing a quick VE fix [18:14:23] Also, wikidatawiki on wmf3? :) [18:14:44] yeah [18:14:49] Thought it was better due to numerous bugfixes [18:14:53] That, and it's a test wiki [18:15:21] Right [18:16:14] OK pushed [18:19:15] !log Run Wikibase/repo/maintenance/createBlacklistedItems.php on wikidatawiki [18:19:24] Logged the message, Master [18:36:10] hah, there's AaronSchulz [18:36:11] :) [18:37:59] thumbs on test2 working now: \o/ http://test2.wikipedia.org/wiki/User:JanGerber/sandbox [18:39:51] Merlissimo: ping [18:40:49] binasher: pong [18:41:30] zhwiki job queue is at ~6mil jobs again, mostly all refreshLinks2 jobs for templates like Country_data_United_States [18:41:44] it sounded like you had idea how to prevent this [18:42:13] last week i created /doc pages for all country data templates. This was requested by Tim [18:42:29] hmm [18:42:57] Tim told be to run a script to remove duplicate jobs, but i think he hasn't done it yet [18:43:25] after my job was finished job queue was about 2.3 mio. [18:43:40] hmm [18:44:13] my bot is currently only updating langlinks on templates if the langlink is included from a subpage and does not edit main templates [18:44:29] Tim-away: ^^ [18:44:39] he won't be around just yet [18:46:11] i'm going to clear a lot of the refreshLink2 jobs and we'll see if they accumulate any differently now [18:48:50] we had a jobqueue problem on dewiki some years ago. Then many templates we rewritten and now we have some people who do only optimize templates on dewiki. maybe you'll find sb. on zhwiki, too [18:49:41] the country data template if horrable implemented because the templatename itself is an argument [18:53:53] AaronSchulz: are you ok if i deploy the memcached-pecl for testwiki change in a bit? [18:54:40] I was about to ask about that, it seems reasonable [18:55:26] binasher: is there any reason why should there be more update jobs than pages? [18:57:33] Merlissimo: the lack of deduplication is one reason [19:01:40] who owns this new "Your edit was saved" thing? [19:01:47] e3? EE? [19:02:18] jeremyb: e3 https://www.mediawiki.org/wiki/Post-edit_feedback [19:02:38] AaronSchulz: how did the refreshLink2 dedupe support work with tim go? [19:02:53] StevenW: ori-l: ping [19:03:08] jeremyb: hey [19:03:08] binasher: I don't think anyone has looked at it so far [19:03:08] Yep, that's us. [19:03:25] StevenW: i'm getting it repeatedly on a page i never edited [19:03:44] oh, i thought you two started hacking on it after the big jobq refactor was merged [19:03:55] load the page again, get the msg again [19:04:06] binasher: besides, this is platform, so getting anyone to work together is like trying to juggle handkerchiefs in a wind tunnel [19:04:10] also got it when loading the history page [19:04:22] haven't dug too much more yet. have you guys heard of that? [19:04:30] jeremyb: can you clear your cookies and see if the problem persists? [19:04:58] ori-l: well sure... but that might throw away the evidence ;P /me will try in a second firefox with it's own cookies [19:05:06] this happened for a few people as a result of a cookie's lifecycle spanning two deployed versions of the extension that handled the cookie differently [19:05:40] ah, interesting [19:05:42] also, what browser? [19:05:56] may not matter, but helps to know [19:06:07] Build identifier: Mozilla/5.0 (X11; Linux i686; rv:14.0) Gecko/20100101 Firefox/14.0.1 Iceweasel/14.0.1 [19:06:12] thx [19:14:45] Can someone kick gerrit where the sun never shines? [19:14:50] Its dead again. [19:15:08] jeblad_WMDE: working for me ? [19:15:33] okey, back again [19:15:36] :) [19:15:58] StevenW: WFM on new cookie jar, still broken on original Iceweasel [19:16:18] jeblad_WMDE: especially cross ocean, i would expect some instability in routing today due to storm stuffs [19:16:26] at&t has already had issues [19:16:26] :( [19:17:07] The cables is on the ocean floor [19:17:12] Doubt they will suffer [19:17:23] But to bad with Bounty [19:17:45] yes the cables are … however they go on land to connect to all of the routers [19:17:51] which also need power [19:17:52] and local cabling :p [19:17:55] jeremyb, ori-l : related… https://en.wikipedia.org/w/index.php?title=Wikipedia%3AVillage_pump_%28technical%29&diff=520456965&oldid=520443330 [19:18:19] jeblad_WMDE: there are some major datacenter buildings in zones b+c at least. and some in zone a i think [19:18:28] have to check the map [19:18:38] of course they have plans and generators [19:19:05] but the storm surge is *already* worse than irene and the storms barely even here yet [19:19:05] "The database has been automatically locked while the slave database servers catch up to the master" [19:19:06] yeah, but sometimes they discover their UPS bank has 5 year old batteries …. [19:19:08] eep! [19:19:46] !log Deleting php-1.20wmf11 files [19:19:59] Logged the message, Master [19:20:13] jeremyb: Fukishima had plans and generators too :-\ [19:20:13] jeremyb: which wiki is this? have you made other edits on that wiki over the past week or so? if so, did you a) see a confirmation, and b) did this problem occur? [19:20:46] ori-l: enwiki. yes i did and i saw a confirmation. didn't try loading so many times before [19:21:01] i tried now because it was a page i'd never edited [19:21:11] actually lets see my contribs [19:21:46] yes, i edited last night [19:22:35] yeah, there's a major cogent DC in zone A [19:24:04] 111 8th ave/76 9th (the building google bought) is zone c [19:24:29] there's datacenters across the street from it which are in zone b [19:26:29] i still can't believe 111 8th was bought by google [19:26:35] i heard 60 hudson is on generator power now [19:27:31] LeslieCarr: it's pretty damn huge. it has it's own hospital! [19:27:40] it does ? [19:27:43] yes [19:27:57] i only know that the Internet (yep, capital I ) flows through it [19:27:59] cool [19:28:09] LeslieCarr: have you been? [19:28:14] nope [19:28:23] only virtually [19:28:28] and the routers don't have eyes [19:28:38] unless skynet is secretly more advanced than I knew .... [19:28:39] hrm [19:29:28] LeslieCarr: there's some really huge stairwells and some smaller ones. one of the larger ones has various things on the walls. 2ish floors are occupied by an xkcd. (the one with the logarithmic scale of what lives at what levels in the atmosphere) [19:29:54] ori-l: any cookies i should look at? or how do i know what they mean? [19:30:23] jeremyb: anything w/'postEdit' in the cookie name [19:31:11] enwikishowPostEdit=1 [19:32:12] AaronSchulz: robla wants the pecl-memcached test to be schedule, so i put it on the deploy calendar for tomorrow with a 1:30-2:30pm window [19:33:48] woosters: ^^ [19:37:09] binasher: we might have played a game of telephone on that one, but it's probably fine that you have a window for it [19:37:53] at first, I forgot the caveat that this was only test2, and so my kneejerk reaction was "yeah, that'll leave a mark" ;-) [19:38:36] I think where we left things after we discussed it on Friday was that it's fine to deploy, just make sure you aren't stepping on someone else's window [19:38:50] (fine to deploy without explicitly scheduling/announcing) [19:41:08] nom time [19:46:05] Reedy: why does MW extensions have no 1.21 milestone on bugzilla? [19:46:27] * Krenair pokes andre__ [19:46:42] Because no one has added any? [19:47:54] Reedy: maybe there was some other reason :) [19:48:44] ori-l: so... do we right this off as "between versions" ? [19:48:59] ori-l: i know for sure that last night was my only edit since i last logged in [19:50:07] jeremyb: i suspect so, but i'm going to investigate further first [19:56:38] errrr, gah. s/right/write/ [19:56:40] ;-( [20:47:35] jeremyb: ping pong ::) [20:47:53] jeremyb: you got some python logic madness pending in a change https://gerrit.wikimedia.org/r/#/c/8120/ [20:47:56] jeremyb: that is the gerrit hook stuff :/ [20:48:31] the inline comments are on PS6 https://gerrit.wikimedia.org/r/#/c/8120/6/files/gerrit/hooks/hookhelper.py,unified [20:48:50] hashar: i noticed. not sure what the current state is. will look later with the puppet refactor [20:48:55] the change has been rebased on top of the gerrit hook test rewrite/enhancement [20:49:03] okkkk [20:49:18] still hoping to get my gerrit hook changes merged one day :-] [21:25:36] error [21:25:42] What did you commit? getting php source [21:25:50] who broke it? [21:25:56] I didn't [21:26:12] Status isn't all good right now [21:26:30] what's going on ? [21:26:39] hahahaha [21:26:40] I got a php page [21:26:48] ?php # WARNING: This file is publically viewable on the web. Do not put private data here [21:26:54] now it works.. [21:27:00] LeslieCarr: For a moment we got source instead of the output [21:27:01] well at least there was no private data there ;) [21:31:30] I just got a more or less raw PHP dump from a Commons page, as if the < was left off the ?php statement at the beginning. I can't tell which server IP address I retrieved the page from and it appears to have been transient and I can't reproduce it, but I have the source code from when it happened. Anyone want any more info about that? [21:32:07] Btw, the api is also affected (I noticed it because my bot started to complain) [21:32:20] * sumanah looks for sbernardin [21:32:27] C-apple-W: i think all we want to know is whether or not it keeps happening [21:33:27] jeremyb: It only happened once to me; the same page came up just fine after that: http://commons.wikimedia.org/wiki/Category:Photographs_by_Pete_Souza [21:34:11] C-apple-W: ok. so if it happens any further then please let us know. ;) [21:35:19] The PHP I got may have come from the Internet-facing servers; it has the individual names of Squid proxies and a bunch of other stuff in the code and contains "# This file is the main and first configuration file of the WMF cluster. # It is included by LocalSettings.php" as lines 6 and 7. [21:36:18] notpeter: hi. can you look at https://bugzilla.wikimedia.org/show_bug.cgi?id=31697 , please? I've told that you may know something about this. [21:37:18] jeremyb: OK, will do. [21:38:05] aharoni: yep, I took a look at it this morning. I'm trying something today to see if it'll fix the problem [21:38:16] thank you! [21:38:25] aharoni: although eventually I'll need a list of affected wikis, for the record [21:38:28] C-apple-W: you weren't the only one, I don't think we need the content. just let us know if it recurs. (or if it changes to a different problem then say so) [21:38:40] but first must actually see if my fix works [21:39:59] notpeter: I can get you that. Do you need something to start with? Do you apply it only to particular languages? Or do you need it for testing? [21:40:29] aharoni: so, the bug report mentions eswiki [21:40:34] and as a californian, I can at least parse spanish :) [21:40:47] if I can solve spanish, we'll go from there [21:40:47] yes, definitely, and that's easy for you for testing :) [21:41:04] but this will involve rebuilding the indexes from scratch, so I'd prefer to not do every single wiki [21:41:05] the reporter came from Hebrew, which I speak. [21:41:10] yeah [21:41:22] I will probably ask for your help on that :) [21:41:31] jeremyb: Ah, OK. Shouldn't be hard to pin down, if it's actually a missing < - if worse comes to worst, find out which file contains that comment, then grep that file on every Wikimedia server until you find ^\?php or whose first 4 characters are ? p h p (which, theoretically, should never occur in a valid file, but I guess you knew that) [21:41:52] C-apple-W: We found it already and fixed it [21:41:52] notpeter: Sure, and ask Asaf and Ori, too ;) [21:41:59] But that discussion happened in another channel [21:42:11] RoanKattouw: OK, excellent! Thanks. [21:42:24] aharoni: excellent! [21:45:31] C-apple-W: i think we're way ahead of you... i think it was only broken for ~30-120 secs approximately. so should be long fixed. so hence i'm asking you to tell us if you see it again [21:45:44] OK, thanks! [21:46:07] C-apple-W: there were a few other people that reported the same thing. if it was more than that 1-2 mins then *a lot* more people would have reported it [21:47:14] OK, I'll let you know if I see it again. Getting off this webchat thing in the meantime-- web-based IRC weirds me out. ;-) [21:47:29] heh, thanks for letting us know [23:29:57] binasher: still nobody has added a plot for zhwiki replag, hasn't it? [23:31:13] zhwiki replag? [23:31:57] äh, jobqueue size [23:32:01] sorry [23:32:27] as it is available for enwiki [23:32:51] i don't think so [23:33:33] is it so difficult? [23:33:48] i think that would help much [23:34:29] http://ganglia.wikimedia.org/latest/graph_all_periods.php?c=Miscellaneous%20pmtpa&h=spence.wikimedia.org&v=506&m=enwiki_JobQueue_length [23:34:57] rewritting it for zhwiki shouldn't be to complicated, i tihnk. But i do not know the code [23:52:51] ori-l: are you re-deploying EventLogging right about now? [23:53:36] chrismcmahon: getting ready to, yeah. would you like to talk it through first? [23:54:46] ori-l: I'm pretty clear on the four patchsets that got reverted, I just want to check some IE8 functions on test2/test on the way to prod, just to be sure [23:55:29] chrismcmahon: sure; i'll ping you once i've updated test / test2 but before i've re-enabled on enwiki. [23:56:11] ori-l: perfect, let me know (and I'll watch IRC) [23:59:16] * andre__ pokes back Krenair in case he's around